Hi Kacper,

thanks, fixed in SVN 823.

Actually, I was thinking of allowing complex numbers for and .

Unfortunately there are different ways for comparing complex numbers, like:

1. first the real part then the imaginary part, or
2. first the magnitude and then angle

But it seems like only 1. is a continuous extension of the real case.

Any opinions?

/// Jürgen


On 12/16/2016 10:55 AM, Kacper Gutowski wrote:
Hello,
There is another case of not handling near-real numbers properly:

      ⍋,1J0
DOMAIN ERROR
      ⍋,1
      ^^

It should of course return ,1.

-k



Reply via email to